# About the lexicon Linear A was the syllabary of Minoan Crete, recorded from about 1800 to 1450 BC on tablets, roundels, sealings, stone libation vessels, metal, and pottery. It precedes Linear B, the script adapted for the Mycenaean Greeks, deciphered through the efforts of many scholars like Alice Kober and Michael Ventris. Linear B took many of its signs from Linear A, and scholars have long agreed these signs kept their sound values. We present an open-source decipherment of Linear A as a lexicon spanning Semitic, Hurrian, Middle Egyptian, Sumerian, Indo-European, and Aegean layers. ## The decipherment Linear A, in our reading, records Minoan, an archaic Semitic language and a sister to Northwest Semitic and Arabic. It builds upon the work of Cyrus Gordon and his successors, who first argued for a Semitic Linear A. Tom di Mino presented the case in [*Ya Diktu: Grammar of the Minoan Peak Sanctuary Libation Formula*](https://www.minoanmystery.org/linear-a/ya-diktu-grammar-tom-di-mino.pdf) (2026). The entire research record is open in [kaphtor](https://github.com/plasma-ai/kaphtor), the source of this site, where each reading stands beside its evidence, its tier and its rivals. The readings are di Mino’s unless an entry names an earlier proposer, as with Gordon’s KU-RO *kull* ‘total’. - [637 words in the lexicon]() - [371 names of people, places, gods and titles]() - [1,719 documents of the corpus]() - [82 inscriptions carved again in 3D]() - [345 Proto-Semitic roots the readings rest on]() - [348 works cited]() ## How far a reading holds Every reading is a proposal, and its tier says how far the evidence carries it. The tier rests on the word’s place in the texts, not on the comparison alone. - [Confirmed]() (18 entries): The function is fixed inside the corpus (position, arithmetic, commodity signs). The etymology may still be a proposal. - [Probable]() (67 entries): Fits the word’s contexts and has a close external comparison, without an independent control. - [Candidate]() (242 entries): A comparison worth testing; the evidence is thin or the word is rare. - [Speculative]() (310 entries): A proposed reading on a word too rare to test. Most occur only once. ## Reading an entry - Signs: The word in Linear A, sign by sign, each sign linked to every entry that uses it. - Reading: The word as it was spoken, vocalized from the signs and the comparison. Linear A writes no final consonants and no consonant clusters, so the reading restores what the script leaves out. - Meaning: The sense the reading gives the word in its contexts. - Confidence: The reading’s tier, as set out above. - Root: The Proto-Semitic root the reading comes from, with its reflexes branch by branch. - Kind: What the word names: a person, a place, a god, a commodity, a vessel, an office, a ritual act and so on. - Language: The layer the reading belongs to: Semitic, Hurrian, Egyptian, Anatolian and others. A layer is a hypothesis about the word, not about the language of Linear A: names and loanwords travel. - Compared forms: The cognates in the related languages, each with its source. A form marked cf. is a parallel, not a cognate. - First proposed: The scholar who first put the reading forward, where it is not di Mino. - Attestations: Every document the word stands on, with its line and the translation’s gloss. - Commentary: The case for the reading: its contexts, its spelling and its parallels. - Rival readings: The other readings on record, named so they can be weighed. - References: The works the entry cites, each linked to the bibliography. ## Signs and abbreviations - `KU-RO`, `RA₂`, `*301`: Linear A signs in capitals, joined by hyphens within a word. A subscript number marks a second sign with the same sound value; an asterisk and a number name a sign by its GORILA number when it has no agreed value. - `*kʷall-`: An asterisk before a form: a reconstruction, attested in no text. - `√k-l-l`: A Semitic root, written as its consonants. - `< >`: Comes from; becomes. - `~`: Varies with; written either way. - `ʾ`, `ʿ`, `ḥ`, `ṣ`, `ṭ`, `ś`, `š`, `ṯ`: The usual transliteration of the Semitic consonants: glottal stop, ʿayin, and the emphatic, lateral and interdental series. - `Heb., Ug., Akk., Aram., Arab., Syr., Phoen., Sum.`: Hebrew, Ugaritic, Akkadian, Aramaic, Arabic, Syriac, Phoenician, Sumerian. - `PS, NWS, LB`: Proto-Semitic; Northwest Semitic; Linear B. - `PN, DN, GN`: Personal name; divine name; geographical name. - `1cs, 3ms, 3fs, 1cp`: Person, gender and number of a verb or pronoun: first person common singular, third masculine singular, and so on. - `G, D, N, tG, Gt stems`: The Semitic verb stems: the base stem, the doubled stem, the n-stem and the stems with an infixed or prefixed t. - `cf., s.v., var., ibid.`: Compare; under the word; a variant; in the same work. - `DULAT, HALOT, CAD, KTU, GORILA`: The dictionaries and editions cited by their sigla, each spelled out in the bibliography. ## The data Our corpus follows the standard edition, GORILA, by Louis Godart and Jean-Pierre Olivier, and its supplement RILA-S1, alongside [SigLA](https://sigla.phis.me), by Ester Salgarella and Simon Castellan, and [lineara.xyz](https://github.com/mwenge/lineara.xyz). We checked its readings against GORILA, John Younger’s *Linear A Texts* and the primary editions, and corrected many of them. The London Medical Papyrus also preserves spells of the Keftiu, the Cretans, written in Egyptian script. Their words remain in kaphtor, but this site keeps to Linear A. The 3D inscriptions are templates we drew at Plasma AI from each text and the object that carries it, not museum scans. This edition of the data is kaphtor [`50525edd`](https://github.com/plasma-ai/kaphtor/tree/50525edd). ## Licence and citation The lexicon, names, translations, commentary and roots are all free to reuse with credit, under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). The corpus keeps the terms of its sources: SigLA’s records are under [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/), and kaphtor’s corrections and added layers under CC BY 4.0. The site’s code is under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0). Cite it as: di Mino, Tom. 2026. Linear A Lexicon. Plasma AI. https://linear-a.ai/. Data: kaphtor, commit 50525edd. ```bibtex @misc{dimino2026lexicon, author = {di Mino, Tom}, title = {Linear A Lexicon}, year = {2026}, publisher = {Plasma AI}, url = {https://linear-a.ai/}, note = {Data: kaphtor, commit 50525edd} } ``` To cite a single reading, give its entry’s address and the work it names. ## Corrections We welcome corrections, whether a wrong form, a missed attestation or a better reading. [Open an issue on kaphtor](https://github.com/plasma-ai/kaphtor/issues) naming the entry and your source, or write to [info@plasma.ai](mailto:info@plasma.ai). Corrections are made in kaphtor first, and reach this site with the next edition of the data. All contributors will be on record and credited in the decipherment. Designed at [Plasma AI](https://www.plasma.ai) with love by Maxim Kuznetsov, Vittorio Retrivi and Tom di Mino. --- Page: https://linear-a.ai/about/ Licence: CC BY 4.0, https://creativecommons.org/licenses/by/4.0/ Cite as: di Mino, Tom. 2026. Linear A Lexicon. Plasma AI. https://linear-a.ai/. Data: kaphtor, commit 50525edd.